Sovereign AI infrastructure for African enterprises.
Corvux deploys open-weight large language models on-premise - giving Kenyan and East African organisations full control over their AI stack, their data, and their compliance posture under the Kenya Data Protection Act (KDPA 2019).
No third-party inferenceand no data crossing borders.
| Layer | What it means |
|---|---|
| Local LLM deployment | Open-weight models running on your hardware |
| Inference infrastructure | Optimised serving stacks - vLLM, Ollama, llama.cpp - tuned for constrained environments |
| Enterprise integration | REST/OpenAI-compatible APIs that drop into existing workflows |
| KDPA compliance tooling | Audit trails, data residency guarantees, access controls |
Cloud-hosted LLM APIs route sensitive prompts - customer data, legal documents, financial records - through foreign servers. For regulated industries in Kenya (finance, health, government), that's a compliance liability and a strategic risk.
Corvux keeps the model, the data, and the inference pipeline inside your perimeter.
Actively building.
We're a small team out of Nairobi. Repositories here are a mix of internal tooling, reference deployment configs, and research.
Reach out via GitHub Discussions or find us on X — @BrianMMwirigi.
